Go back to John Quiggin's home page
Does privatisation pay ?
Centre for Economic Policy Research,
Australian National University
Department of Economics
James Cook University
Quiggin, J. (1995), 'Does privatisation pay ?', Australian Economic Review 95(2), 23-42.
I would like to thank Peter Forsyth, Clive Hamilton, Max Neutze, Bob Walker and participants in a University of New England seminar for helpful comments and criticism.
Privatisation has become an popular policy since 1980, particularly for governments facing difficulty in raising revenue. However, there has been no consistent attempt to estimate the impact of privatisation on the net fiscal position of the public sector, or even to consider how such an estimate might be made. The object of this paper is to consider the implications of privatisation from a fiscal perspective.
Privatisation automatically reduces budget deficits in the short and medium term. However, this apparent benefit reflects the generally defective nature of the budget deficit as a measure of public saving.
In this paper, case studies of a number of actual and proposed privatisations are presented. In each case it is shown that the savings in public debt interest associated with privatisation are insufficient to offset the loss to the public sector of the earnings of the enterprise concerned. In many cases, the sale price is around 50 per cent of the present value of the stream of earnings foregone. This suggests that the loss in public sector net worth is as large, or larger, than the sale price.
This article shows that the public debt interest savings associated with privatisation are, on average, less than the profits foregone, implying that privatisation reduces public sector net worth.
Does privatisation pay ?
Privatisation has become an popular policy since 1980, particularly for governments facing difficulty in raising revenue. The British government alone has raised over £ 60 bn stg by this method. Privatisation is widely recommended as a method of restoring fiscal rectitude. On the other hand, socialist and other critics of privatisation have condemned it as 'selling the family silver to pay the bills'. To complicate the picture further, a number of economists have argued that privatisation will have essentially no impact on the government's fiscal position.
There is surprisingly little evidence to support or refute any of these views. There has been no consistent attempt to estimate the impact of privatisation on the net fiscal position of the public sector, or even to consider how such an estimate might be made. Despite the vast amounts of money involved, privatisation programs have been adopted (and opposed) on the basis of faith rather than of evidence.
The object of this paper is to consider the implications of privatisation from a fiscal perspective. Issues of efficiency and equity will be considered primarily as they affect the fiscal analysis. Thus, for example, in considering the issue of improvements in productivity efficiency associated with privatisation our concern will be how much of this improvement can be captured by government through privatisation. Similarly, the question of whether the apparent benefits of privatisation represent transfers from public employees or pure efficiency gains will be addressed only briefly.
The primary focus is on the policy of using the proceeds of asset sales to pay off debt. This is the preferred policy of most serious advocates of privatisation. However we will briefly examine two alternatives. The first is the use of privatisation to fund current spending and the second is the allocation of privatisation proceeds to new equity investments.
We will consider a number of case studies. Some of these relate to actual privatisations -- including British Telecom, the Commonwealth Bank and NZ Telecom. The others relate to proposed privatisations which did not in fact go ahead at the time and involve a reassessment of proposals to privatise the Commonwealth Bank in the mid-eighties, made by Coughlin (1987) and the large-scale privatisation proposed in the Fightback! manifesto (Hewson and Fischer 1991).
The naive budget framework
In terms of the standard budgetary accounting conventions, the case for privatisation of government business enterprises (GBEs) appears unanswerable. The whole of the sale price of the asset concerned is recorded as revenue in the year of sale (or, in the British case, as an offset against outlays). In addition, assuming the proceeds are used to pay off debts, public debt interest (PDI) is reduced in every subsequent year. Against this, the only offset is the loss of dividends remitted to the government.
This analysis reflects the generally defective nature of the budget deficit as a measure of public saving. The budget deficit is
(i) simply a measure of cash flow and
(ii) is confined to the 'budget sector'.
The fact that the budget is a cash flow measure means that capital and current expenditures are lumped together. Similarly asset sales from within the budget sector are treated as current income (or, in some budget systems, as an offset against current expenditure). Thus, the budget deficit is useful as a measure of public saving only if the level of net public investment (new investment less asset sales) is essentially constant. For this reason, it is common to adjust reported budget deficits to exclude the revenue from asset sales.
Further difficulties arise from the division of the public sector into budget and non-budget components. GBEs lie outside the budget sector, and are disregarded altogether in the budget statement unless they are sold or remit dividend payments to the government. Transfers between the budget and nonbudget sectors are essentially irrelevant in measuring public savings. This fact is not widely appreciated. For example, in the Fightback! package, the Labor government's practice of treating the proceeds of asset sales as current revenue was severely (and correctly) criticised. But the same package treated an accelerated repayment of Telecom debt to the government ( that is, a transfer from the nonbudget to the budget sector) as an expenditure saving.
In an economic analysis, the appropriate comparison is between the flow of savings in PDI arising from privatisation, and the flow of earnings (including capital gains) that would have accrued in the absence of privatisation (adjusted for any differences in tax treatment etc.). The impact of the sale price on revenue in the year of sale should, of course, be disregarded altogether. Furthermore, the dividend policy imposed on the enterprise is also essentially irrelevant.
The main conceptual difficulty is the treatment of inflation. Two equivalent solutions to this problem may be offered. The first is to use the nominal bond rate to discount the savings in PDI and to include nominal capital gains in the flow of income accruing under continued public ownership. The alternative is to use the real bond rate for discounting and count only real capital gains. The latter approach is more appealing, especially since it will frequently be necessary to disregard unrealised capital gains and focus only on measured net earnings.
The real bond rate cannot be observed directly since it depends on inflationary expectations. Mehra and Prescott (1985) estimate that over the present century, the long-run average real rate has been around 1 per cent. However, it seems likely that at least part of this reflects unanticipated inflation in the post-war period. At least for the foreseeable future it seems unlikely that such low rates will prevail. We will use a rate of 5 per cent, but will undertake appropriate sensitivity testing.
Privatisation and fiscal illusion
In this paper, attention will be confined to the issue of whether privatisation per se increases the net worth of the public sector. The implicit assumption is that privatisation decisions have no impact on spending and taxation decisions. In effect, the proceeds of privatisation are used entirely to reduce government debt.
However, this assumption may not be valid in practice. At least some government decision-makers appear to accept the naive budget framework in which the proceeds of privatisation are regarded as current income. For example, the proposed sale of the Commonwealth Airports Corporation has been justified, at least in part, as a method of financing increased expenditure on labour market programs. The Thatcher government financed large tax cuts out of the proceeds of privatisation. Since, as will be argued below, these privatisations actually reduced public sector net worth, the effect was to exacerbate public dissaving. The current fiscal crisis in the UK may be traced in large measure to the fiscal illusion associated with privatisation.
Even when the trap of counting privatisation proceeds as current income is avoided, it is rare to find a correct understanding of public sector savings. For example, the Fightback! package (Hewson and Fischer 1991) avoided the use of current-year privatisation as a source of finance, and proposed using the proceeds to repay debt. However, the Fightback! calculations treated the whole of the reduction in public debt interest as a net gain, with no allowance for the earnings foregone (not even the component remitted as dividends to the public sector).
In summary, even if privatisation can be undertaken without a reduction in public sector net worth, it may have the effect of reducing the rate of public savings. Whether this is desirable in any particular case will depend on a great many circumstances, but it is generally preferable that decisions on public savings should be made in a transparent fashion, rather than being clouded by fiscal illusions.
Arguments for and against equivalence
A number of authors, notably Forsyth (1994), have argued that privatisation should have no effect, or very little effect, on public sector net worth. This equivalence hypothesis may be supported as follows. Assuming that the profits of the enterprise are unaffected by privatisation and that investors discount returns at the real government bond rate, the sale price of the enterprise will be exactly equal to the expected present value of the flow of profits it generates. Unfortunately, neither of these assumptions is accurate. Moreover, since the first assumption is typically violated in a way that tends to raise the attainable sale price and the second in a way that tends to reduce it, it is impossible a priori to determine the direction of bias.
The claim that privatisation permits improvements in profitability is at the core of the case for privatisation. The question of whether privatisation per se is associated with profit gains is discussed in the following section. From a fiscal viewpoint what matters is the extent to which any profit gains can be captured in the sale price. If any private operator can realise improvements, the government can sell the GBE for the present value of the flow of profits under private ownership. On the other hand, if there is only one or only a few potential buyers who can realise increased profits, the rent will be shared between the government and the purchaser. Nevertheless, to the extent that there are profit gains, the sale price will tend to be higher than the value of the stream of future profits under continued public ownership, discounted at the private discount rate.
The equity premium
The assumption that private buyers will discount the expected flow of returns at the bond rate (adjusted for an appropriate risk premium) is based on an apparently simple arbitrage argument. Investors can choose between bonds and stocks as devices to smooth lifetime consumption flows and any premium in the required rate of return for stocks can therefore be explained by the level of relative risk aversion. Unfortunately the observed premium is much too large to be explained in this way. This is the equity premium puzzle of Mehra and Prescott (1985).
The result is that the discount rate used by stockmarkets in evaluating the expected flow of returns from a privatised GBE may be much higher than the government's opportunity cost of funds. Hence, other things being equal, a policy of selling profitable assets and using the proceeds to pay off debt will reduce public sector net worth (Walker 1994).
The issue of which rate of return is appropriate depends, in large measure, on the way in which the equity premium puzzle is resolved. If the high premium can be derived as the result of the optimising choices of a rational representative consumer (as suggested, for example, by Constantinides 1990), then in some sense, the true opportunity cost of public investment in GBEs is the equity rate of return. However, it is difficult to see how such an argument could be constructed without, at a minimum, assuming Ricardian equivalence, and the empirical evidence for Ricardian equivalence is very weak indeed.
The existence of an equity premium appears to provide a strong case for public sector holdings of equity, at least up to the point where the level of public sector indebtedness begins to impose costs, for example, by generating a risk premium for government bonds. However, it may be argued (Forsyth 1994) that this does not affect the case for privatisation. Forsyth presents two related arguments.
First, he argues that if governments could really exploit the difference between the equity and debt rates it would amount to a 'Pactolus' or river of gold, that governments could exploit to obtain a limitless flow of returns. This argument is incorrect, since the expansion of equity investment would ultimately equate the returns to equity and debt. Further limits are imposed by the fact that in industries, such as agriculture, where efficient operation is inconsistent with the separation of ownership and control, the organization of production by a government business enterprise (or indeed, any widely held corporate structure) would not be feasible.
Forsyth varies the River of gold argument to suggest that, if expanded public holdings of equity were profitable, governments would already be exploiting them. This simply begs the question. Many governments have undertaken large-scale privatisation in the belief that major reductions in public equity holdings are desirable. Other governments have steadily expanded their net holdings of equity. The question of what the optimal level actually is cannot be decided by appeal to the observed actions of government.
Forsyth also suggests that instead of using privatisation proceeds to repay debt, governments could undertake alternative equity investments (active or passive). This argument is valid, but it is important to realize its limits. It does not weaken the case against a policy of selling assets (or otherwise reducing public equity holdings) and using the proceeds to pay off debt. It merely states that, assuming the public sector holding of equity is already optimal, the public sector should be willing to sell low yielding assets if they can be replaced with higher-yielding assets. Obviously, no general statement can be made about the desirability or otherwise of transactions in which one asset is sold and another is bought. However, in situations where a large equity premium indicates that existing public equity holdings are too small, it would normally be preferable to retain the first asset and fund the purchase of the second by increased debt.
The possibility, suggested by Forsyth, of using passive equity investments as a buffer against fluctuations in the stock of capital in the government business enterprise sector, deserves further exploration. However, the public sector capital stock is large in relation to the stock of privately owned equity capital. Hence, if the proceeds of a systematic privatisation campaign were applied to equity investment, the public sector would end up as either a majority shareholder or a substantial minority shareholder in a large number of companies. In neither case could the notion of passive investment be sustained. The difficulties associated with, for example, takeover bids and more general conflicts of interest would be similar to those associated with existing government business enterprises.
The equivalence argument is also affected by the political incentives involved in pricing an enterprise for flotation (or, in the case of sale by tender, in deciding whether to accept the best offer). Particularly in situations where privatisation is a controversial policy, there are substantial political costs in a float that is undersubscribed or a tender process that ends in failure. There are also political costs in underpricing an asset, with the result that share purchasers make immediate capital gains, but these may be offset by the concentrated political support for privatisation that will be generated among the shareholders. In the case of sale by tender, no clear basis for a claim of underpricing can be made, whereas a failure to generate an acceptable bid cannot be hidden. Thus, in a situation where the market value of the enterprise is unclear, the political calculus favors underpricing.
Evidence from the UK suggests that underpricing has been the rule, particularly in the case of public flotation. Most of the public floats offered in the eighties were oversubscribed (by factors ranging from 1.5 to 34, with a typical oversubscription rate of about 5 to 1) and prices in open market trading exceeded the offer price from the first day of trading, on average by about 25 per cent of the full price (Vickers and Yarrow 1989). Since the shares were offered on a partly paid basis, investors who sold their shares as soon they were issued frequently made capital gains of 60 per cent or more. (The main exception was BP, where the October 87 stock market crash took place while the offer was under way. As a result the float was severely undersubscribed).
Australian evidence similarly favours this conclusion. The initial float of shares in the Commonwealth bank was at a price that yielded substantial gains to purchasers. This was sufficiently widely recognised that the shares were heavily oversubscribed. When the second tranche of shares was sold, the market price was observable and the offer price was very close to the market price (since the shares subsequently declined in value, purchasers made capital losses).
The moral hazard problem
Another difficulty relates to the possibility of future changes in policy. It is possible, as suggested by Walters (1989) in relation to the relatively low price obtained for British Telecom, that investors are concerned by the possibility of renationalisation. It is not clear what losses investors would in fact incur under renationalisation. Clearly the possibility of nationalisation without compensation could be ignored, so that the main risk would be that shares would be compulsorily acquired at the original offer price, or at some price based on the offer price rather than the ruling market price. A more serious problem arises in the case where firms are privatised but continue to be regulated. There now exists a moral hazard problem, in that governments have an incentive to tighten regulations in response to favorable shocks leading to high profits, while forcing shareholders to bear the brunt of unfavorable shocks. Although governments may make more or less credible commitments not to behave in this way, it appears that moral hazard problems of this kind are, to some extent, an inescapable consequence of privatisation. As such they must be set against the various agency problems associated with public ownership. The problem of policy changes is discussed in more detail by Zeckhauser and Horn (1989).
More generally, the problem of allocating risk between governments and private suppliers of infrastructure and other services is one which has not been resolved in Australia. Particularly in pseudo-private infrastructure projects such as toll roads, the private partner faces sufficient risk to demand a high rate of return on capital, but the government still faces a large and undefined residual risk. The result is that the project is far more costly to taxpayers than if it were constructed on a turnkey basis and financed by bond issues with the repayments being met by user charges.
The converse problem arises with the sale of enterprises providing essential services such as water and telecommunications. The government can not avoid the necessity of regulating and the possibility of adverse regulatory decisions means that private bidders offer low prices for the assets. Yet the risk faced by the government is not significantly reduced by privatisation since the enterprises are too important to be permitted to fail. Thus the taxpayers face the worst of both worlds. Assets with high expected earnings are sold for low prices, but the government is left with large contingent risk exposures.
As long as privatisation and pseudo-private infrastructure projects are characterised by an undefined allocation of risk between the government and the private owner, the moral hazard problem will ensure that the price received is less than the value of the flow of earnings foregone, appropriately adjusted for the change in the government's risk exposure.
In evaluating any large-scale policy program, the central problem is the appropriate choice of counterfactual. The first class of counterfactuals consists of estimates of the realisable value under privatisation for firms that have in fact remained in public ownership. Although such estimates have not been prepared on a routine basis a number of estimates have been prepared, in most cases, in the context of proposals for privatisation. For example, a central element of the Fightback! manifesto presented by the Liberal party as its platform for the 1993 election campaign was a large-scale privatisation program, estimated to yield $20 billion. Similarly Coughlin (1987), advocating the privatisation of the Commonwealth Bank presented calculations to support a (1984-85) market value of $1.5 billion.
The advantage of treating this kind of example, is that actual profits in continued public ownership can be observed. Furthermore, since the estimates of market value have normally been prepared by advocates of the privatisation proposal in question, they may assumed to be free of any downward bias. Since, in the cases considered here, it may be shown that the value in public ownership exceeds the calculated sale value, the possibility of upward bias in estimated market values is not a critical difficulty.
The main difficulty is that examples of this kind are relatively hard to find. Furthermore, there is somewhat more direct interest in determining whether privatisations that have taken place have been profitable than in determining whether those that have not taken place would have been profitable. Logically, however, these issues are two sides of the same coin.
The difficulties of choosing an appropriate counterfactual are greater in the case of privatisations that have gone ahead. It is possible to observe reported flows of profits before and after privatisation, but not the profits that would have been made in the absence of privatisation. In valuing an asset, a natural assumption to consider is that the stream of earnings will continue at the same level as at the time of sale. However, in the case of privatisation, this assumption is problematic. The central difficulty is that privatisation occurs in the context of a program of reforms aimed at increasing the profitability of GBEs and reducing emphasis on other objectives, such as the provision of secure and well-paid employment to workers and the supply of goods and services at low prices to consumers judged to be deserving. Privatisation has generally taken place part of the way through this process, and has therefore occurred in a context of rising profits.
One view of this process is that privatisation is an essential element such that, without the expectation of privatisation, little or no reform would be feasible (Domberger and Piggott 1986). On this view, the appropriate counterfactual is the continuation of the level of profits (or, as in many cases, losses) achieved prior to the commencement of reform .
At the opposite extreme it may be argued earnings without privatisation would have been the same as those actually observed under privatisation. In support of this view may be advanced the contention, made in a number of studies of privatisation, that it is market structure, rather than ownership that is the primary determinant of productive efficiency. It follows that governments could mimic the effects of the privatisations that have actually taken place.
Neither of these extreme views seems plausible. In regard to the first, it may be useful to consider in detail the range of reforms that may be considered. First, there is the reform of prices to remove subsidies to users. Second, there is reform of work practices to increase labor productivity. Third, there is the withdrawal of above-market wages and conditions for employees. Fourth, there is shedding of excess labour and the closure of loss-making sections of the enterprise. Fifth, there is the removal of constraints on capital investment associated with limits on public sector borrowing. Sixth, there is the restructuring of markets to permit competition. Finally, there is the introduction of new products and management techniques permitting flexible responses to changes in market conditions. The first six of these appear to be quite consistent with continued government ownership, with only the seventh creating obvious difficulties
Although pricing reform involves political difficulties, examples of such reforms in publicly owned enterprises are easy to find. For example, when the Whitlam government abolished the old Postmaster-General's Department to form Australia Post and Telecom Australia as statutory corporations it introduced a general increase in prices aimed at securing sufficient net profit to permit both internal financing of new investment and the payment of a return on capital. Subsequent price changes were to be reviewed by the Prices Justification Tribunal. This once-off reform proved sufficient to achieve a permanent return to profitability.
Reform of work practices, wages and conditions
As an example of reforms aimed at increasing labour productivity is the reorganisation of the US Postal Service in 1970. Prior to this reorganisation, labour productivity (measured in prices of mail per employee) had been virtually static. Subsequently, it rose by about 3.3 per cent per year over the period 1970-83 (Statistical Abstract of the US, cited in Peltzman 1989).
Although the reorganisation reduced the losses suffered by the Postal Service, much of the benefit went to Post Office employees, whose wages increased relative to those of the manufacturing workforce as a whole by about 2 per cent a year over the post-reorganisation period (during most of which real wages were declining for manufacturing as a whole). It has frequently been argued that government ownership necessarily weakens the bargaining position of the public employer vis-a-vis employee unions. It is certainly true that many publicly owned firms have provided better wages and conditions and, more frequently, greater security of employment than comparable private firms. However, this would appear to be a reflection of the political objectives of the governments concerned, and the greater freedom from market constraints associated with public ownership.
In cases where governments have been concerned to push wages down, public ownership has similarly facilitated the task. In such periods, the wages of public sector employees have typically fallen relative to those in the private sector. Similar considerations apply in relation to industrial disputes. Governments that wish to avoid such disputes will typically settle on easier terms than private employers. On the other hand, where governments want to break union power, the facts that they are not concerned with bankruptcy and that they can act simultaneously as a party and as a referee in the dispute, gives them significant advantages relative to private firms. The mobilisation of state power against the National Union of Miners by the Thatcher government and against the air traffic controllers union (PATCO) by the Reagan government was greatly enhanced by public ownership.
Probably the biggest single element of the reform program of the eighties and nineties has been labour-shedding. This reflects the unwillingness of governments in the seventies, when high and rising unemployment was widely seen as a temporary aberration, to announce large-scale redundancies, particularly in areas where unemployment was already above-average. Another aspect of public enterprises tending to reduce labour-shedding was the existence of almost unconditional tenure of employment in the core civil service and its partial extension to public enterprises, particularly those that had formerly been government departments. It was widely argued in the seventies and early eighties that large-scale labour shedding (usually referred to in this context by some euphemistic term such as 'flexibility in employment levels') was not possible for public enterprises. This argument appears to have been invalidated by events. Although considerable labour-shedding has typically been undertaken in firms being prepared for privatisation, the process has been even more rapid in enterprises, such as the railway industry, where privatisation is, at best, a possibility for the remote future.
The case of Telecom Australia is also of interest in considering this issue. Telecom's employment has fallen by more than 20 per cent over the past five years . The present government has made commitments not to privatise Telecom, at least until 1997, though the history of such commitments reduces their credibility. If the analysis linking reform to privatisation is to be applied in cases of this kind, then, it is apparent that only the possibility of privatisation and not an overt commitment, is required.
Another interesting feature of the Telecom case is that almost all of the management team that commenced the program of reform will have left the enterprise by the end of 1994. This raises the question of how the possibility of privatisation can generate incentives for managers to behave differently than if they expected continued public ownership. In the case of imminent privatisation, the incentive is presumably the possibility of obtaining highly paid management positions (and possibly some element of management control) in the privatised firm. But this incentive cannot apply over the medium term. Furthermore, it appears to be strongly applicable only in the case of sale by public flotation. Sales by tender would be more likely to generate the replacement of the existing management team by appointees of the purchaser, and the likelihood of a such a sale would probably reduce performance incentives for incumbent managers.
Constraints on capital investment
The need for capital beyond public sector borrowing constraints has often been advanced as a motive for privatisation. Prima facie, this is the weakest of all the arguments for privatisation that have been put forward. The public sector borrowing constraints are themselves, in large measure, an innovation of the reform period. In large measure this innovation is based on hypotheses such as 'crowding out'. The crowding out model (which is by no means universally accepted) is based on an assumption that government debt is used to finance current consumption. Even if the crowding out model is accepted, the appropriate aggregate for targeting is the difference between revenue and current consumption expenditure.
All that is required, then, is to free selected public enterprises from the constraints, and permit them to issue their own debt. The main objection to this is the suggestion that all such debt will be implicitly guaranteed by the full faith and credit of the government. This is obviously a matter of degree. Public enterprise or 'semi-government' debt has historically attracted a risk premium over pure public debt, although this premium has not been as great as that for similar private enterprises. On the other hand, private firms providing crucial services have frequently been regarded as 'too important to fail' and have been rescued by governments, usually on terms that provide at least some benefit to bondholders. Hence, the critical policy issue is that governments should make explicit their willingness or unwillingness to guarantee the debts of public or private enterprises.
Market restructuring and competition
Public enterprises have generally operated in markets where competitive entry has been strictly limited, and many analysts have taken this to be a defining feature. In the UK case, the main exceptions have been where ailing private firms have been rescued by government. As a result, the claim that the competitiveness of the market structure rather than the ownership of the firm is important has been met with the response that privatisation is a necessary, if not a sufficient condition for competitive market restructuring. It is argued that the government can never be trusted to act fairly in regulation if it owns one of the contending firms.
Australian experience does not support this contention. Australia has perhaps been unique in the extent to which GBEs have operated side-by-side with private firms, in industries such as airlines, insurance, banking, radio and television and, more recently, telecommunications. In most cases, the private operators in these industries have been regulated and there have been some restrictions on entry. However, extensive deregulation of entry has been undertaken in circumstances where privatisation was either not contemplated, or at least a relatively distant eventuality. In banking for example, the Commonwealth Bank was established in 1913 as a competitor for the private banks (State banks were established even earlier). State governments also established insurance companies, serving markets such as third-party motor vehicle insurance that had, in some cases been abandoned by private insurers in the face of rising payouts, and also competing with the private sector in the provision of life insurance, household insurance and other products. Extensive financial deregulation was introduced between 1981 and 1985 at a time when privatisation of the Commonwealth Bank and State government financial enterprises had not been seriously advocated. Similarly, domestic airline deregulation was introduced in the late eighties. Although privatisation of the publicly owned airlines was contemplated at that time, and is likely to take place, the airlines remained in majority public ownership in 1994. Finally, in the case of Telecom Australia, the government deliberately introduced competition to a market previously served by a monopoly GBE. The terms were considerably less favorable to the incumbent than those which applied, for example, to the privatised British Telecom. In particular, all long-distance subscribers were balloted to make a choice between Telecom and the entrant firm Optus. Optus was entitled, in the event of inadequate participation in the first ballot, to call for a second, and has indicated that it will exercise this right.
Although competitive deregulation will probably be followed by privatisation in most cases of this kind, this simply reflects the association between the two ideas. The government could, if it chose, halt privatisation now (as the Queensland state government has done in the case of its financial enterprise Suncorp) and continue to hold majority ownership in enterprises in each of the now deregulated industries discussed above.
The claim that privatisation is a necessary precondition for reform appears strongest with respect to the introduction of new products and flexible response to changing market conditions. GBEs have frequently been capable of substantial technical innovation (before privatisation, British Telecom was responsible for innovations such as Teletext, which prefigured the 'information superhighway' frequently discussed today). However, they have generally not been adept at responding to market conditions. The procedures of public administration are geared to the implementation of relatively stable rules, with emphasis on adherence to correct processes. To some degree or other, almost all GBEs are affected by these emphases. Hence they are characterised by less precise discrimination between markets and relatively slow responses to changing conditions.
In summary, though, the claim that the prospect of privatisation is necessary to promote reform, or even a major stimulus to reform, does not seem tenable. The great majority of the reforms that have affected GBEs destined for privatisation may be observed in cases where privatisation was at best, a distant possibility. The association between reform and privatisation is primarily one of political compatibility rather than of logical implication.
Should we, then, adopt the alternative of assuming that the profits foregone by government are those actually realised under privatisation? There are several arguments against this. Most obviously, private enterprise firms are explicitly oriented towards the maximisation of profit. It would be surprising if any alternative structure could outperform them in achieving this goal. The argument for a complete reform under public ownership essentially envisages governments acting like the shareholders of a private company. But shareholders must rely either on direct supervision of the companies they own or on the exit option of selling their shares on the open market. In the case of GBEs, direct supervision must imply a significant element of political control over management decisions with all of the associated difficulties. The exit option is, of course, that of privatisation.
Moreover, although renationalisation is a possibility, privatisation has much more finality than a program of reform in the context of continued public ownership. A change of government, or simply of political sentiment, could lead to the reassertion of other objectives at the expense of profit maximisation. Hence, even if a government enterprise was made to work in a purely profit-maximising fashion, it would be difficult to maintain this indefinitely.
In this section, case studies of a number of actual and proposed privatisation are presented. The central issue in each case is whether the savings in public debt interest associated with privatisation (assuming real rates of 5-6 per cent) are sufficient to offset the loss to the public sector of the earnings of the enterprise concerned. The answer is negative in each case.
British Telecom 1985
The privatisation of British Telecom in 1985 was a pivotal event in the history of privatisation. Before 1985, privatisation in Britain was essentially a continuation of the denationalisation policies that all incoming Conservative governments had pursued. A variety of enterprises that had been brought into public ownership by Labor administrations, either through nationalisation or through the rescue of firms in difficulty, were sold back to the private sector. The most important policy that did not fit into this framework was the sale (usually to existing tenants at subsidised prices) of public housing. Although this policy was radical in the British context, it is commonplace in other countries, including Australia.
British Telecom was different. Telecommunications had developed under public ownership and was part of the standard portfolio of government business enterprises in most developed countries (with the noteworthy exception of the US). Standard theories of public economics provided a rationale in terms of natural monopoly that appeared to fit telephone services perfectly. Hence the decision to privatise BT represented a radical break with the past, and a move to a stance in which government would seek to withdraw entirely from the production and sale of goods and services.
The sale of BT also introduced a new regulatory method for monopolies -- the RPI-X pricing rule according to which the movement in the prices of a defined basket of services should fall short of the general movement in retail price levels by at least X per cent. The RPI-X rule was proposed by Littlechild (1981). Evaluation of this method of regulation is outside the scope of the present paper. However, it may be observed that the rule imposed on BT, in which X was set at 3 per cent, was very generous, and calculated to maximise the sale price rather than to impose substantial pressure of the privatised firm to improve performance. As Beesley (1992), a strong advocate of privatisation, observes
'BT's past productivity record indicated to the government an X of -2.5 per cent at the least: much more than -2.5 might jeopardise the proceeds from privatization. The compromise reached was an X of -3' (negative signs in original).
Given the rate of technical progress in telecommunications, the chosen X of 3 per cent guaranteed that the privatised firm could enjoy increasing profits without any internal reform and that the benefits of any reform that did take place would flow entirely to shareholders. Indeed, the report advocating the use of the RPI-X approach to regulation indicated that it under RPI-X, BT's prospects, and the associated sale price would be much the same as if BT was sold without any explicit constraints on pricing and profitability.
BT was sold in two stages. Attention is here focused on the first stage sale of 50 per cent of the shares. The offer price of shares was 1.30 stg, implying a valuation of the entire firm at 7.8 b stg, in comparison to estimated assets of 16 b stg in 1981-82 (British Treasury, cited by Beesley and Littlechild 1983). Shares were offered on a partly paid basis with an initial payment of 50p and the partly paid shares traded at 93p on the first day, indicating significant underpricing.
In addition, small shareholders received incentives such as bonus shares. According to the National Audit Office (cited by Mayer and Meadowcroft 1986), these incentives, along with other sale costs, reduced the net return to government from the sale of British Telecom by 263m or about 5 per cent of the gross proceeds of the sale. Thus the net proceeds from the sale of the first 50 per cent of BT were about 3.65bn stg.
In 1984/85 BT had a gross operating surplus of about 3000m and interest liabilities of 500m. (BT prospectus and HM Treasury government expenditure plans). Assuming that a commercialised BT would face an effective tax rate of 40 per cent, this would imply a net profit of 1.5bn, or 750m for the half-share sold in 1984-85.
This implies that the effective rate of return foregone by the first-stage privatisation was 20.5 per cent, as against a real bond rate of around 5 per cent. Alternatively, taking the present value of the stream of profits, evaluated at the bond rate, the valuation of the stream of profits is around 15bn as against a net sale price of 3.6. If an 8 per cent rate is used the valuation becomes 9.5bn. That is, the partial privatisation of BT reduced the net worth of the British public sector by at least 5 bn and probably 10 bn.
Part of the 20.5 per cent rate may be accounted for by the costs of sale and by underpricing. At the first days trading price, the private half-share of BT was valued at 5.1bn implying a rate of return of around 15 per cent. The predicted dividend yield was about 4 per cent. Although the 15 per cent rate of return (implying a price-earnings ratio below 7) is a little high relative to other stocks, the predicted dividend yield rate is quite low. This suggests that the market anticipated that profits would remain broadly in line with the 1984/85 values.
The offer of a sale price below the value of measured assets is a consistent feature of the privatisations and proposed privatisations that have been examined here. Except in declining industries such as coal and steel, the value of assets would normally represent a minimum commercial value. In the case of BT, a firm in a growing industry privatised with essentially intact monopoly status, and under very generous conditions, the market value would be expected to exceed the value of capital assets.
However, this analysis encounters difficulties associated with the equity premium problem. If the firm consistently follows a policy of undertaking investments whenever they yield a rate of return equal to the real bond rate, its marginal investments will reduce its present value evaluated at the discount rate for equities. Hence, it is possible that the market value of a nationalised firm may be below its asset value.
Privatisation in the UK - an overview
Although the sale of British Telecom was the pivotal event in the process of privatisation in the UK, it was small in relation to the total program. Is there any way of evaluating the overall fiscal impact of privatisation in the UK?
One approach would be to begin with an assessment of the fiscal impact of the GBE sector as a whole before privatisation commenced. Pryke (1986) argues that, after account is taken of depreciation and related factors, the GBE sector yielded either zero or a small negative profit in 1979. So, if 1979 was taken as representative, it might be argued that the 60bn stg realised from privatisation represented a net fiscal gain.
There are several difficulties with this argument. First, some of the big loss-makers of 1979, notably the coal industry and British Rail, remained in public ownership in 1994, though their losses had been greatly reduced. With their removal from the analysis, it appears likely that the GBE sector would have been profitable in 1979.
Second, the profits of the GBE sector were in fact increased greatly (and could have been increased further ) by reform programs with no necessary link to privatisation. Indeed, as already observed, the biggest gains were realised prior to privatisation.
Third, 1979 was an atypical year. The economic downturn commencing in 1974 and the associated wage-price spiral had reduced the profit share of GDP. GBEs were under additional pressures. Attempts to control inflation led to downward pressure on prices, while the emergence of mass unemployment for the first time in nearly 40 years made it difficult for loss-making firms to reduce staff numbers through processes of natural attrition or negotiated redundancy agreements. In these circumstances, large losses were accepted as an inevitable concomitant of the times. Even in the absence of any program of GBE reform, it is unlikely that this situation would have continued indefinitely. GBEs would have shared in the general recovery in the profit share. Upward adjustments in prices would have been permitted. Once the permanent nature of the unemployment crisis had been recognised, some form of compulsory redundancy (or closure) would have been imposed on loss-making firms.
The Commonwealth Bank 1985
Australia is one of the few OECD countries with substantial government ownership of banks. In the context of financial deregulation, it was scarcely surprising that proponents of privatisation should turn their attention to the Commonwealth Bank. Coughlin (1987) presented a case for privatisation. He concluded that the Bank could have been sold in 1985 for about $1.6 bn (or $2.5 billion 1993/94) as against a book value of assets of $1.9 bn. Coughlin valued the expected stream of dividends under public ownership at about $800m and concluded that the bank should be sold.
Coughlin's approach is similar in spirit to that advocated here. However, his analysis suffers from two serious errors. First, in common with the Budget accounting procedure he focuses on the flow of dividends, rather than the flow of earnings. By the Modigliani-Miller theorem, in the absence of differential tax treatment (which is clearly irrelevant to the Commonwealth government), the owners of a firm should be indifferent as regards its dividend policy. Indeed, under Coughlin's approach, the value of the Bank in public ownership could be increased massively by the simple book-keeping measure of notionally paying out all earnings as dividends and having the government reinvest as much of the earnings as desired.
Second, Coughlin uses nominal discount rates, but takes no account of nominal capital gains. As argued above, the appropriate procedure is to use a real discount rate in the range from 1 per cent to 8 per cent, with 5 per cent as the preferred value.
Profits for the Commonwealth Bank are given in Column 1 of Table 1. For 1983-84 and 1984-85, the Bank was not fully liable for company tax and Coughlin proposes an adjustment to estimate the post-tax profit under standard tax treatment. Allowing for double taxation of dividends under the classical system, which prevailed until the late eighties, some further adjustments are required. If it is assumed (following Coughlin) that 35 per cent of earnings would normally be remitted as dividends and taxed at 30 per cent earnings must be adjusted downwards by 10.5 per cent. This correction is given in Column 2. In Column 3 the values derived in Column 2 are converted to 1993-94 dollars.
Commonwealth Bank profitability
FY end Net profits Profits (adj) Real profits
1984 202 170 281.2
1985 228 221 343.2
1986 277 277 400.3
1987 198 198 265.0
1988 359 359 449.6
1989 476 476 550.1
1990 524 524 569.3
1991 883 883 921.3
1992 409 409 419.7
1993 443 443 450.0
1994 682 682 682.0
We may now consider the valuation of the Bank in public ownership. Assuming 1985 earnings would have continued in perpetuity yields an estimate constructed on a basis similar to that used by Coughlin after the correction of the errors in his valuation procedure. Discounting 1985 earnings at 5 per cent would imply a valuation of $5.6 billion, more than double the market valuation suggested by Coughlin. Even using an 8 per cent rate, privatisation would have implied a large loss for the government.
In fact, however, profits increased significantly in real terms after 1985. If we take the average profit over the period 1985-94 as representative of the future income stream, the value of the bank in public ownership rises to $10 billion. Thus, Coughlin's proposal for privatisation would have implied a loss to the public sector of 75 per cent. Indeed, no projections of future profits are required. Using a 5 per cent real rate of discount, the stream of profits earned over the period 1985-94 alone has a net value of $3.9 billion -- more than the entire market price estimated by Coughlin.
It may of course be argued that this improvement was unforeseeable at the time Coughlin's estimate of the sale price was made. However, it was already evident that deregulation had greatly improved the competitive position of the major banks relative to their domestic competitors (particularly the building societies) and that the threat of foreign competition had been greatly overrated. Hence a significant increase in profitability was likely.
The Commonwealth Bank. 1991
When Coughlin's proposal for privatisation was advanced, the ALP represented itself as an implacable opponent of privatisation, and this opposition was used as the basis of several successful election campaigns. By 1991, however, the situation had changed sufficiently to permit the sale of a large tranche of shares in the Commonwealth Bank.
The capital structure established for the bank prior to the sale involved the issue of 835 million shares. Although the par value for the shares was set at $2, the relevant consideration for valuation is the issue price which was set at $5.40. This implies a valuation of $4.5 billion for the Bank as a whole (or $4.7 billion 1993-94 dollars). A second tranche of shares was sold in 1993. The sale procedure in this case ensured that the government would receive an amount very close to the market price of the shares at the date of sale, which turned out to be around $9.50, implying a valuation for the entire bank of $7.9 billion. Share prices subsequently rose above $10.00 before falling to around $8.00.
Using the earnings set out in Table 1, and assuming for the moment that the 1991 issue price represented the value of the shares in public ownership in 1991, we may now consider the effect of undertaking privatisation in 1991, rather than in 1985 as suggested by Coughlin. As already observed, privatisation in 1985 would have yielded a return of $2.4 billion (93-94 dollars). Delaying privatisation until 1991 yielded a substantial increase in the value of the enterprise, in addition to the flow of dividends returned to the Budget sector over this time. The use of the 1993 market values makes the 1985 privatisation proposal look even less attractive. Indeed, the revenue obtained from the sale of 29 per cent of the shares in 1993 was equivalent (in real terms) to the value placed by Coughlin on the entire bank in 1985. The government has already realised more from a partial privatisation as it would have from Coughlin's complete privatisation proposal, while still retaining ownership of an income stream with a present value greater again than the sale price proposed by Coughlin.
However, the comparison between earlier and later privatisation is of essentially historical question. A more relevant issue is whether the current privatisation process is financially beneficial to the Commonwealth. Average real profits over the period 1988-93 (which covers a complete business cycle) have been $560 million. Computing the present value this stream of profits at a discount rate of 5 per cent yields a value of $11.2 billion for the bank as a whole or $3.25 billion for the 29 per cent of shares actually sold, as against sale proceeds of $2.3 billion. Hence the partial privatisation of the Commonwealth Bank reduced the net worth of the public sector by approximately $1 billion. Using an 8 per cent discount rate, the second tranche privatisation was a marginally profitable proposition for the government.
In summary, the fact that privatisation was undertaken in 1991 rather than at the time of the first serious privatisation proposal in 1985 nearly doubled the return to the public sector and the price increase for the 1993 tranche yielded a further 80 per cent improvement. Despite this, even the 1993 share sale resulted in a reduction in public sector net worth at standard discount rates.
Telecom New Zealand
The case of New Zealand Telecom is of interest because sale to foreign telecommunications firms was preferred to the more common vehicle of a public float. From the analysis above, it would be expected that such an approach would be likely to yield a higher sale price. On the other hand, the consequences of any underpricing are more severe, since losses to the public sector are not offset by gains to the domestic private sector.
Prior to 1988, telecommunications services in New Zealand were provided by the NZ Post Office which was organised as a government department. New Zealand Telecom was organised as a publicly owned corporation in 1988. Over the period 1988 to 1990 annual profits were around $NZ250m. In September 1990, the entire corporation was sold to a consortium of US telecommunications companies for a price of $NZ4.25 billion. The regulatory regime was very liberal, with heavy reliance being placed on the introduction of a competitor, CLEAR, also backed by American firms. At the time of privatisation, the government predicted rapid increases in profits. This prediction was borne out, with profits reaching $NZ420m in 1992.
Based on the counterfactual assumption that profits in the absence of privatisation would have continued at the level prevailing before privatisation, the sale price is roughly equal to the present value of profits foregone, suggesting that there was no loss of public sector net worth associated with privatisation.
However, the 'no change' counterfactual appears rather too generous in this case, for two main reasons. First, the period between corporatisation and privatisation was very brief. The general success of corporatisation in increasing profits suggests that significant profit increases could have been realised under public ownership. Second, the combination of rapid technical progress with a natural monopoly in local services gives the telecommunications industry great potential for profit growth whenever regulation does not directly constrain profits. An assessment of the privatisation of Telecom NZ based on the assumption that the profits stream foregone was the same as that actually observed implies that public sector net worth was reduced.
Nevertheless, a comparison between the privatisation of NZ Telecom and that of British Telecom suggests that the New Zealand taxpayers got a much better deal. The British government sold half of BT for a net price of about stg3 bn, or, allowing for the exchange rate and inflation, about NZ10bn, implying a valuation of the entire enterprise at $NZ20bn. Considering that the British market is more than ten times as large as the New Zealand market, and that BT was sold with considerable monopoly privileges attached, the price received for BT looks very low indeed.
One of the most significant, but comparatively little discussed, features of the Fightback! package was the proposal for privatisation of up to 600 Commonwealth GBEs with an estimated total sale price of $20 billion over four years. The savings in public debt interest associated with privatisation were the biggest single expenditure saving items proposed in the package. As has already been noted, no allowance was made for the loss of dividend income to the Budget sector or, more generally, for the loss of GBE profits to the public sector as a whole.
Privatisation was even more important for the financing of the Fightback Mark II package, released in December 1992. In this package, the proceeds from privatisation were allocated to the funding of large-scale capital works expenditure rather than debt reduction. As has been argued above, such a shift is sound in principle, though the actual program included a number of measures which were current rather than capital in nature and also included some capital expenditure programs which could not be justified in cost-benefit terms, or which would not yield returns to the public sector, and hence would ultimately reduce public sector net worth..
Unfortunately, Fightback! did not provide a detailed analysis of the assets proposed to be sold. However, the first year program (assumed to apply from 1992-93) included the sale of remaining shares in Qantas and the Commonwealth Bank, along with the AIDC, Snowy Mountains Engineering Corporation and Commonwealth Serum Laboratories, with estimated returns of $5 billion. The AOTC (now Telstra) was to be sold in three tranches, commencing in the second year of the program. The AOTC sale would have been by far the biggest single element of the privatisation program for these three years, which involved estimated returns of $15 billion. The total program would yield real savings in public debt interest of $1 billion/year assuming a 5 per cent real interest rate.
Consideration of Telstra alone indicates that this saving would be more than offset by the loss of GBE earnings. The profits of Telstra and its predecessors Telecom Australia and OTC are shown in Table 2. For the years 1988-89 and 1989-90, when Telecom did not pay tax, a tax rate of 40 per cent has been assumed.
Table 2 - Telstra net profits
Telecom OTC Telstra
1988-89 973* 189 772.8
1989-90 1288* 233 1005.8
1990-91 963 275 1238
1993-94 1700 (est)
* - Not taxed
Table 2 shows that, averaging over the last six years, the loss of the profits of Telstra alone negates the benefits in reduced public debt interest of the entire privatisation program. In other words, the fiscal effect of the Fightback! program would have been to give the other 599 enterprises away for no return to the public sector.
The Fightback! program looks even worse if we disregard 1992, when profits were reduced by extraordinary losses associated with the Telecom-OTC merger, and suppose that profits can be maintained at or near the 1993-94 level. On this assumption, Telstra alone generates profits that are nearly double the total savings in public debt interest of the entire privatisation program. In present value terms, Fightback! proposed to sell an asset with a value of at least $30 billion for less than $15 billion and then throw hundreds of others into the bargain.
Looking only at the first-year program, in which Telstra was not involved, the results of the previous section indicate that the return of $5 billion (implying an annual real PDI saving of $300 million) would not have been sufficient to offset the loss of the 80 per cent interest in the Commonwealth Bank, taking no account of the other enterprises proposed for sale.
In summary, the privatisation program proposed in Fightback! would have greatly reduced the net worth of the public sector. This would ultimately have necessitated either higher levels of taxation or expenditure cuts deeper than those proposed in Fightback! Even if privatisation ultimately proceeds, the delay in implementation associated with the electoral rejection of Fightback! will be highly beneficial. Taxpayers would benefit even more if the entire notion of large-scale privatisation were permanently abandoned.
Fiscal benefits and social benefits
This paper has focused on the question of whether privatisation yields financial benefits to governments (and therefore, ultimately, to taxpayers) and has derived a generally negative answer. Nevertheless, it is perfectly possibly for a policy to be costly to taxpayers but nonetheless socially beneficial. It is important, therefore to consider whether there are benefits to other groups that would yield a case for privatisation.
It is evident that privatisation has rarely been beneficial for the employees of former government business enterprises, with the exception of top management. Of course, it is necessary to observe that much of the losses in wages, conditions and employment levels that has taken place in privatised firms would have taken place anyway in the context of microeconomic reform, so that the impact of privatisation per se has been small. Nevertheless, what impact there is has undoubtedly been negative.
Nest, we may consider benefits to consumers. Privatisation has generally been associated with the removal of various cross-subsidies, benefiting some consumers and harming others, but with a general presumption of an increase in consumer welfare. Privatisation has also been accompanied by some increases in productive efficiency. However, as with the effects on employees, this process is largely independent of privatisation per se. In relation to the cases considered above, it is reasonable to suggest that the privatisation of the Commonwealth Bank has had almost no effect on consumers and that the failure to privatise earlier also had no effect. In relation to Telstra, the effect on consumers would have depended on the nature of the market environment in which privatisation took place, and this was not spelt out clearly enough in Fightback! to permit an evaluation of consumer effects. It seems unlikely that privatisation would have led to improvements in productive efficiency more rapid than those that have taken place under public ownership.
In the case of British Telecom, it seems clear that privatisation was associated with a reduction in consumer welfare, although some of this was gradually recouped through the reimposition of regulation. Not only was the RPI-X regulation extremely generous, but the initial regulatory structure failed to take account of the fact that BT had a strong incentive to reduce service quality in those areas where it had a clear monopoly, and particularly in the case of loss-making CSOs such as the provision of public telephones.
Finally, we may consider gains to shareholders in the privatised firms and to participants in financial markets. It seems clear that these groups have benefitted significantly. In particular, the underpricing of shares in initial floats is essentially a transfer from taxpayers to those allocated shares. It is tempting to assume that such transfers are costless. This is incorrect, for several reasons.
First, in some cases, the beneficiaries of the transfer are wealthy foreigners, whose interests are not normally taken into account in evaluating policy outcomes. In the case of British Telecom, 11 per cent of shares went to foreigners (Mayer and Meadowcroft). Similarly, although the underpricing of Telecom NZ was less than for any of the other cases examined here, the fact that the benefits went entirely to foreign corporations makes the social costs of privatisation larger.
Second, raising taxes is socially costly. Most estimates suggest that the deadweight cost associated with an additional dollars taxation is around 20 cents, although some estimates (ironically usually from advocates of privatisation) suggest losses around 50 cents per marginal dollar. Since the transfers to shareholders associated with privatisation rarely have any redeeming equity effects and since the loss of value to the public sector must ultimately necessitate increased taxation, it is reasonable to regard a substantial portion of any such transfer as a deadweight loss.
Third, privatisation has been accompanied by increased financial sector activity, with sale costs typically equalling between 2 and 4 per cent of the sale price and with subsequent share trading implying additional use of resources in the financial sector. Although this has undoubtedly contributed to bidding up the returns to those engaged in financial market activity, the increase in financial sector resources used in asset markets must be regarded as a pure social loss.
In summary, because privatisation is essentially a financial transaction, the effects of privatisation in reducing the net worth of the public sector are likely to dominate the social welfare effects of the program. Although the loss in public sector net wealth will be partially offset by the transfer to shareholders, such transfers are both costly and inequitable.
The equity premium and the welfare analysis of privatisation
Thus, much of the transfer from taxpayers to shareholders and financial market participants associated with privatisation must be regarded as a genuine social loss. However, the bulk of the loss computed here is derived from the equity premium, that is, the difference between the marginal cost of funds to government and the post-tax rate of return demanded by holders of equity.
Although a number of explanations have been advanced for the equity premium none has yet gained general assent. For the purposes of the presented paper, two classes of explanations may be distinguished. In the first class of explanation, the equity premium is derived as the result of optimising decisions of consumers in perfect capital markets with preferences consistent with expected utility and In the second class, the premium is derived either from a violation of the expected utility hypothesis (Epstein and Zin 1990) or from a market failure in the form of non-diversifiable risk faced by individuals.
In general, the optimising solutions are likely to imply that the government's attitude to holding equity should be similar to that of private individuals and hence that discounting at the bond rate is in appropriate. Even if these solutions to the equity premium are accepted, however, it would appear sensible for the government to hold some assets in the form of equity. Moreover, there are substantial difficulties in the way of an optimising solution to the equity premium problem and, in particular, to the application of such solutions to public finance. First, all of these solutions involve very strong assumptions about the capacities of individuals for rational optimisation over very long(in some cases, infinite) time-horizons. In particular, any application to public finance implies the validity of the Ricardian equivalence hypothesis (Barro 1974) and empirical support for this hypothesis is very limited. Second, the most plausible hypothesis of this kind, advanced by Constantinides, requires a peculiar form of intertemporal nonseparability in preferences, implying a strong aversion to short-term variations in consumption along with a willingness to accept large long-term variations. Although this may be a plausible description of the behavior of some individuals, it does not appear compelling as a guide to policy when individuals are conceived of as representatives of infinitely-lived dynasties, as is required for the validity of the Ricardian approach.
If a non-optimising, or non-expected utility, solution to the equity premium puzzle is ultimately derived, the case for large public holdings of equity and for the use of something close to the bond rate for purposes of discounting becomes very strong. The case for public holding of equity comes down to the ability of the state to pool risks through the mechanism of taxation, thereby circumventing the market failures that generate the equity premium. On this analysis, the bond rate should be used for discounting as long as the risks associated with public holdings in equity are small in comparison to the capacity of the government to raise taxation revenue. On this point, it may be observed that even the State banking disasters of the eighties cost, in total, less than 5 per cent of one years revenue for the Australian public sector as a whole. It seems clear that any pure risk premium associated with public sector equity holdings of the current magnitude must be small compared to the observed equity premium.
Fiscal necessity has frequently been urged as a justification for privatisation. However, the examples studied in this paper suggest that, in general, privatisation serves to reduce the net worth of the public sector. In many cases the sale price is around 50 per cent of the present value of the stream of earnings foregone. This suggests that the loss in public sector net worth is as large, or larger, than the sale price. In terms of the impact on public sector finances, the most favorable (or rather, least unfavorable) of the privatisations considered here was the sale of New Zealand Telecom. On the fairly generous assumptions of a 6 per cent discount rate and no increase in profits in the absence of privatisation, this sale could be regarded as fiscally neutral. If it is supposed that at least some of the profit increases following privatisation could have been captured under public ownership, then this sale, along with all the other privatisations and proposed privatisations considered here, was fiscally unsound.
The apparent budgetary improvements associated with privatisation reflect the inadequacies of public accounting conventions, rather than any real fiscal improvement. The loss of public sector net worth has, on occasion, been exacerbated by this fiscal illusion, which has led governments to introduce tax cuts, supposedly financed by privatisation proceeds.
Several reasons may be identified for the loss to taxpayers associated with privatisation. First, assets are, on average sold for less than market value because of the political imperatives associated with privatisation. Second, 'moral hazard' problems associated with subsequent regulation of the enterprise may reduce market value. Third, because the rate of return demanded by holders of equity is well above the government's cost of funds, even a sale at full market value will in general leave the government worse off. A final source of loss is the possibility that fiscal illusion associated with reductions in the measured budget deficit will lead to reduced rates of public saving.
The adverse effects of privatisation have been most noticeable in the UK, where prices have been set well below market value in public floats and where the effects of fiscal illusion have been particularly serious. The very poor fiscal position of the UK, with the budget deficit measured at 7.8 per cent of GDP in 1993, may be traced, in large measure, to the effects of privatisation.
Abelson, P. (1987), Privatisation: An Australian Perspective, Australian Professional Publications, Sydney.
Barro, R. (1974), 'Are government bonds net wealth?', Journal of Political Economy 82, 1095-1117.
Beesley, M. (1992), Privatization, regulation and deregulation, Routledge/Institute for Economic Affairs, London.
Beesley, M. and Littlechild, S. (1983), 'Privatisation:Principles, problems and policies', Lloyds Bank Review (149), 111-37.
Constantinides, G. M. (1990), 'Habit Formation: A Resolution of the Equity Premium Puzzle', Journal of Political Economy 98(3), 519-43.
Coughlin, P. (1987), 'The Commonwealth Banking Corporation: A case for privatisation', in Abelson, P. (ed.), Privatisation: An Australian Perspective, Australian Professional Publications, Sydney.
Domberger, S. and Piggott, J. (1986), 'Privatisation policies and public enterprise: A survey', Economic Record 62(173), 145-62.
Epstein, L. G. and Zin, S. E. (1990), ''First-Order' Risk Aversion and the Equity Premium Puzzle', Journal of Monetary Economics 26(3), 387-407.
Forsyth, P. (1994), 'The states, microeconomic reform and the revenue problem', paper presented at Conference of Economics, Gold Coast.
Hewson, J. and Fischer, T. (1991), Fightback ! Taxation and expenditure reform for jobs and growth, Panther Publishing and Printing, Canberra.
Mayer, C. and Meadowcraft, S. (1986), 'Selling public assets: Technical and financial implications', in Kay, J., Mayer, C. and Thompson, D. (ed.), Privatisation and Regulation: The UK Experience, Clarendon Press, Oxford.
Mehra, R. and Prescott, E. C. (1985), 'The Equity Premium: A Puzzle', Journal of Monetary Economics 15(2), 145-61.
Pryke, R. (1986), 'The comparative performance of public and private enterprise', in Kay, J., Mayer, C. and Thompson, D. (ed.), Privatisation and Regulation: The UK Experience, Clarendon Press, Oxford.
Pryke, R. (1986), 'The comparative performance of public and private enterprise', in Kay, J., Mayer, C. and Thompson, D. (ed.), Privatisation and Regulation: The UK Experience, Clarendon Press, Oxford.
Quiggin, J. (1992), 'Free lunches in the case for privatisation', Economic Analysis and Policy 22(1), 1-18.
Vickers, J. and Yarrow, G. (1986), 'Telecommunications: Liberalisation and the privatisation of British Telecom', in Kay, J., Mayer, C. and Thompson, D. (ed.), Privatisation and Regulation: The UK Experience, Clarendon Press, Oxford.
Vickers, J. and Yarrow, G. (1989), 'Privatisation in Britain', in Macavoy, P., Stanbury, W., Yarrow, G. and Zeckhauser, R. (ed.), Privatization and State-Owned Enterprises, Kluwer Academic Publishers, Boston.
Walker, R., (1994), Privatisation: A reassessment, University of NSW.
Walters, A. (1989), 'Comment on Vickers and Yarrow', in Macavoy, P., Stanbury, W., Yarrow, G. and Zeckhauser, R. (ed.), Privatization and State-Owned Enterprises, Kluwer Academic Publishers, Boston.
Zeckhauser, R. and Horn, M. (1989), 'The control and performance of State-owned enterprises', in Macavoy, P., Stanbury, W., Yarrow, G. and Zeckhauser, R. (ed.), Privatization and State-Owned Enterprises, Kluwer Academic Publishers, Boston.
Go back to John Quiggin's home page